Workshop on Media Forensics

Overview

The recent advent of techniques that generate photo-realistic fully synthetic images and videos, and the increasing prevalence of misinformation associated with such fabricated media have raised the level of interest of media forensics within the computer vision community. Both academia and industry have addressed this topic in the past, but only recently, with the emergence of more sophisticated ML and CV techniques, has multimedia forensics has become a broad and prominent area of research. In the recent years, there has been a number of events (Media Forensics@CVPR2019 or WOCM@ECCV2018 among others) that touched on this topic but now, more than ever, it critical to keep advancing on all fronts of media forensics: from detection of manipulations, biometric implications, misrepresentation/spoofing, etc.

Topics

• Media forensics
• DeepFakes
• Detection of forged and manipulated images, documents and video
• Discrimination between computer graphics and photographic images
• Reconstruction of media genealogy, provenance
• Semantic interpretation
• Analysis and detection of imagery and videos created by new synthesis methods such as Generative Adversarial Networks (GANs)
• Counter-forensics, i.e. methods to defeat traditional media forensics
• All aspects of anti-spoofing and deception detection in visual information systems
• Datasets and competitions

Competitions

A special section of the workshop will be devoted to discuss results on some of the top media forensics related competitions. We encourage researchers to use the following datasets to report results and submit papers to any of the following three competitions:

• DeepFakes Detection Challenge (description + Kaggle competition)
• FaceForensics (GitHub)
• ChaLearn Face Anti-spoofing Attack Detection Challenges:
  • Track 1: Chalearn Multi-modal Cross-ethnicity Face anti-spoofing Recognition Challenge
  • Track 2: Chalearn Single-modal (RGB) Cross-ethnicity Face anti-spoofing Recognition Challenge

Sponsors